new Cog VMs available
Locked
 
|
...in http://www.mirandabanda.org/files/Cog/VM/VM.r3000/
CogVM binaries as per VMMaker.oscog-eem.772/r3000 Add libcrypto and libssl to Linux and Mac builds of the SqueakSSL plugin. Add the XDisplayControlPlugin and include it (external) and the AioPlugin (internal) to the linux x86 builds.
Fix an abort (relocating call to invalid address) due to an over-zealous check in relocateCallBeforeReturnPC:by:. Since we relocate e.g. calls to primitives there can be no effective range check there-in.
Spur: Fix fillObj: signedness for objects straddling the midpoint of the address space (quickly affects linux). Similarly for routines in pigCompact, to get asserts correct.
Fix printOopsFrom:to:. for objects up to endOfMemory. Declare lastFreeChunk and firstFreeChunk correctly. Fix numberOfForwarders: and printForwarders: for isForwarded:'s blindness towards freeChunks.
Comment isForwarded: to be clear on the issue. Have the segment manager pass to sqAllocateMemorySegmentOfSize: the address of the first large enough gap in the address space, instead of the address of the
end of the first segment. This allows e.g. linux to use MAP_FIXED and hence get past a 128Mb limit on mmapping. Fix bugs in isValidFreeObject: & printFreeTreeChunk: that caused bogus assert
failures. Remember to include the Spur moniker in the -version output on linux -- best, Eliot
|
|
|
On 14 June 2014 05:59, Eliot Miranda <[hidden email]> wrote: > > ...in http://www.mirandabanda.org/files/Cog/VM/VM.r3000/ > > CogVM binaries as per VMMaker.oscog-eem.772/r3000 Always happy to have to update CI to use the latest Cog :) But one question I asked yesterday, and need an answer to - are you abandoning the YY/WW/rev location path? It's not a biggie if you do; it just means I need to touch a few more lines of code. frank |
Re: [squeak-dev] new Cog VMs available
|
|
In reply to this post by Eliot Miranda-2
On Fri, Jun 13, 2014 at 10:11 PM, Chris Cunnington <[hidden email]> wrote:
Providing they work... Any news?
best, Eliot
|
|
|
In reply to this post by Frank Shearar-3
On 14 June 2014 11:23, Frank Shearar <[hidden email]> wrote: > On 14 June 2014 05:59, Eliot Miranda <[hidden email]> wrote: >> >> ...in http://www.mirandabanda.org/files/Cog/VM/VM.r3000/ >> >> CogVM binaries as per VMMaker.oscog-eem.772/r3000 > > Always happy to have to update CI to use the latest Cog :) > > But one question I asked yesterday, and need an answer to - are you > abandoning the YY/WW/rev location path? It's not a biggie if you do; > it just means I need to touch a few more lines of code. Never mind, false alarm. I forgot that the yy.ww.rev scheme is in the name of the artifact, but the path remains Cog/VM/VM.rNNNN. Sorry for the noise! frank > frank |
Re: [squeak-dev] new Cog VMs available
|
|
In reply to this post by Eliot Miranda-2
The latest Seaside (3.1.1) and Grease (1.1.0) load fine and all tests pass. I loaded it with this command: (ConfigurationOfSeaside3 project version:'3.1.1') load: 'Tests' with the configuration from MCHttpRepository location: 'http://smalltalkhub.com/mc/Seaside/MetacelloConfigurations/main' user: '' password: '' Profiling the tests gives: Spur v3000: 9168ms Squeak 4.5 all-in-one: 13102ms Squeak 4.5 all-in-one image + Cog v3000 vm: 7782 msec. Hope this helps Paul |
Re: [squeak-dev] new Cog VMs available
|
|
That Squeak 4.5 all-in-onr number is a little high. I re-ran it several times and its now closer to 10s, e.g. 10281ms
|
Re: [squeak-dev] new Cog VMs available
|
|
Hi Paul,
re: > Spur v3000: 9168ms > > Squeak 4.5 all-in-one: 13102ms > > Squeak 4.5 all-in-one image + Cog v3000 vm: 7782 msec. What platform are you on? Did you have a web browser open when you ran the timings?
And if you did, if you close that browser and rerun do you get different timings? I say this because using Spur for the Newspeak bootstrap I see Spur being twice as fast as Cog when there's no web browser open, but running at essentially the same speed if there is my usual browser open (which has ~ 30 tabs open and in that state is a cycle hog). I'm concerned that Spur may need to manage its priority and/or i/o rate to obtain the potential performance, which would be a real drag.
On Sat, Jun 14, 2014 at 8:02 AM, Paul DeBruicker <[hidden email]> wrote:
best, Eliot
|
Re: [squeak-dev] new Cog VMs available
|
|
Hi Eliot,
That was on a Mac, with Mavericks. I did have Chrome open with about 7 tabs. Closing the browser and rerunning the tests had no noticeable impact on the outcome. To get the timings I just used the Test Runner's "Run Profiled" command and selected every test loaded by Seaside. On Windows 7 the Spur v3000 vm loads Seaside and runs the tests with no problems. But I can't give you a comparison because I'm getting the attached crash.dmp file when attempting to load the ConfigurationOfSeaside3 using the Cog v3000 vm I'll try it on ubuntu now and report back in a bit. crash.dmp
|
Re: [squeak-dev] new Cog VMs available
|
|
Hi Eliot,
On Ubunutu 13.10, All Seaside 3.1.1 & Grease 1.1.0 tests, no browser open Spur v3000: 11445ms Cog v3000: 9675ms On every platform so far the profiler output just shows that they both spend most of their time updating the progress bar the test runner displays. Hope this helps and let me know what else you want to see. Paul
|
Re: [squeak-dev] new Cog VMs available
|
|
Oh and that was with the coglinuxht and cogspurlinuxht vms if it wasn't apparent.
|
«
Return to Squeak VM
|
1 view|%1 views
| Free forum by Nabble | Edit this page |